Download Open problems on countable dense homogeneity

OPEN PROBLEMS ON COUNTABLE DENSE HOMOGENEITY
MICHAEL HRUŠÁK1 AND JAN VAN MILL2
Abstract. We survey recent development in research on countable dense homogeneity
with special emphasis on open problems.
1. Introduction
A separable space X is countable dense homogeneous (CDH) if, given any two countable
dense subsets D and E of X, there is a homeomorphism f : X → X such that f [D] = E.
This is a classical notion that can be traced back to the works of Cantor [15], Brouwer [13],
and Fréchet [35]. Examples of CDH-spaces are the Euclidean spaces ([13, 35]), the Hilbert
cube ([33]) and the Cantor set. In fact, every strongly locally homogeneous (SLH1) Polish
space is CDH, as was shown by Bessaga and Pelczyński [12] (see also [11, 4, 31, 36]).
The term countable dense homogeneous was coined by Bennett in his seminal paper [11]
where the fundamental themes of the subject were introduced by either proving a theorem,
or by asking an insightful question: its relation with other notions of homogeneity such as
strong local homogeneity and n-homogeneity, issues dealing with products, open subspaces
and connectedness. The subject was subsequemtly popularized and advanced greatly by
the efforts of Fitzpatrick [24, 25, 26, 27, 28, 29, 30].
Several surveys on homogeneity and countable dense homogeneity have been written
[30, 27, 6]. In [27], Fitzpatrick and Zhou presented the most interesting open problems on
countable dense homogeneity at the time. Now we are almost 25 years later in time. Since
there is ongoing interest in the subject, it seems appropriate to see what has remained of
their problems and which new ones have emerged. As Fitzpatrick and Zhou remarked, the
problems range in flavor from the geometric to the set theoretic. This remains very true
of the subject today.
2. Fitzpatrick and Zhou’s Problems
First we recall the problems from [27] and review the progress made.
Problem 1. Is every connected CDH metric space SLH?
Date: December 30, 2016.
1991 Mathematics Subject Classification. 54H05, 03E15, 54E50.
Key words and phrases. countable dense homogeneous.
The second-listed author was supported by a PAPIIT grant IN 102311 and CONACyT grant 177758.
1Recall that a space X is strongly locally homogeneous [32] if it has a basis B of open sets such that for
every U ∈ B and every x, y ∈ U there is a autohomeomorphism h of X such that h(x) = y and h restricts
to the identity on X \ U .
1
2
HRUŠÁK AND VAN MILL
This question was implicitly asked in van Mill [59]. In 1990, there were two known
examples of spaces that are CDH but not SLH. The first one, due to Fitzpatrick and
Zhou [26], is Hausdorff but not regular, and the second one, due to Simon and Watson [89]
is regular but not completely regular. The problem was finally solved by van Mill [62].
He showed that there is a connected, Polish, CDH-space X that is not SLH. In fact, any
homeomorphism of X that is the identity on some non-empty open subset of X is the
identity on all of X.
An important special case of Problem 1 remains unsolved. Ungar [85, 84] proved that
that CDH metric continua are n-homogeneous2 for all n. Kennedy [48] showed that a 2homogeneous metric continuum X must be SLH, provided that X admits a nontrivial homeomorphism that is the identity on some nonempty open set. Whether every 2-homogeneous
metric continuum admits such a homeomorphism remains an open problem.
Problem 1’. Is every CDH metric continuum SLH?
It was claimed by Ungar [85] that every dense open subset of a locally compact separable
metric CDH-space is again CDH. The proof is, however, incomplete. This prompted
Fitzpatrick and Zhou to ask the following:
Problem 2. If X is CDH and metric and U is open in X, must U be CDH?
And more specifically:
Problem 2’. If X is connected, CDH, and metric, and U is an open, connected set in X,
must U be homogeneous? If U is homogeneous, is it necessarily CDH?
Problem 2 and the first part of Problem 2’ were answered by van Mill [63]: there is a
CDH Polish space with a dense connected open subset that is rigid. Here a space is called
rigid if the identity function is the only autohomeomorphism. The question whether every
dense open subset of a locally compact separable metric CDH-space is again CDH remains
unsolved.
There are more problems which deal with connectedness issues.
Problem 3. If X is a CDH, connected, Polish space, must X be locally connected?
This is known to have an affirmative answer in case when X is also locally compact
(Fitzpatrick [24]), but the only examples we know of CDH connected metric spaces are also
locally connected. Partial results were recently obtained by van Mill in [68].
Problem 4. For which 0-dimensional subsets X of R is X ω homogeneous? CDH?
The first part of this problem was answered by Lawrence [51], and further extended by
Dow and Pearl [19] who proved that X ω is homogeneous for every zero-dimensional first
countable space X. See also Motorov [73] and van Engelen [21] for earlier partial results.
a space X we let H (X) denote its group of homeomorphisms. According to Burgess [14], a
space X is n-homogeneous if given two sets F, E ⊆ X each having exactly n-many elements, there is a
homeomorphism h ∈ H (X) such that h[F ] = E. The space X is strongly n-homogeneous if given two
sequences {ai : i < n}, {bi : i < n} of points in X both of length n, there is a homeomorphism h ∈ H (X)
such that h(ai ) = bi for every i < n.
2For
OPEN PROBLEMS
3
The second part of the question is still being investigated though important partial
results were obtained: First Fitzpatrick and Zhou in [28] showed that Qω is not CDH
(here Q denotes the space of rational numbers) and then Hrušák and Zamora-Avilés [44]
extended their result by showing that every separable metric space X such that X ω is
CDH is a Baire space. They solved the problem for Borel zero-dimensional spaces by
showing that a Borel CDH space must be completely metrizable (see also [65, 42]), and,
in particular, for Borel X ⊆ 2ω , the space X ω is CDH if and only if X is Gδ . This raised
the question whether there can be a non-Gδ set X such that X ω is CDH. This problem
stimulated research on topological properties of filters (treated as subspaces of 2ω ) treated
in a series of papers [55, 39, 41, 49, 76] where it was established that for a filter F on ω the
following properties are equivalent: (1) F is CDH, (2) F ω is CDH, and F is a non-meager
P-filter3. Quite recently Medini in [53] constructed a ZFC example of a set X ⊆ 2ω which
is not completely metrizable yet X ω is CDH. It is interesting to note that his example is
consistently analytic, while it is also consistent that there are no analytic examples.
Problem 5. Is the ω th power of the Niemytzki plane homogeneous?
This problem is open.
Problem 6. Does there exist a CDH metric space that is not Polish?
Problem 6’. Is there an absolute example of a CDH metric space of cardinality ω1 ?
Both Problems 6 and 6’ were solved in the affirmative by Farah, Hrušák and Martı́nez
Ranero [23]. They proved that there exists a CDH-subset of R of size ℵ1 . Their example is
a so-called λ-set, i.e., a space in which every countable set is Gδ . Note that every countable
CDH-space is discrete, hence ℵ1 is the first cardinal where anything of CDH-interest can
happen. Since R is CDH and has size c, it is an interesting open problem what can happen
for cardinals greater than ℵ1 but below c. It was shown recently in Hernandez-Gutiérrez,
Hrušák and van Mill [40] that for every cardinal κ such that ω1 ≤ κ ≤ b4 there exists a
CDH subset of R of size κ. More about this in the next section.
To recapitulate: Problems 1, 2, 2’ (first part), 4 (first part), 6 and 6’ are solved, while
parts of problems 2’, and 4’, and problems 1’, 3, and 5 remain open. We shall return to
some of these questions in more detail in what follows.
3. New list of some new and some old problems
3.1. CDH and related notions of homogeneity. A topological space X is homogeneous5
if for every x, y ∈ X there is an h ∈ H (X) such that h(x) = y. Every CDH space is a
that a filter F is a P-filter if given a sequence {Fn : n ∈ ω} of elements of F there is an F ∈ F
such that Fn \ F is finite for every n ∈ ω.
It is a major open problem in set-theory whether non-meager P-filters exist [45]. They do exist in all
known models of set theory, and their possible non-existence implies the existence of measurable cardinals.
4Recall that b = min{|F | : F ⊆ ω ω ∀g ∈ ω ω ∃f ∈ F ∃∞ n ∈ ω f (n) > g(n)} is a combinatorially
defined cardinal number whose value may be any regular cardinal between ω1 and c.
5The notion was introduced by Sierpiński in [80].
3Recall
4
HRUŠÁK AND VAN MILL
topological sum of clopen homogeneous subspaces, in particular, every connected CDH
space is homogeneous [11, 25]. It was noted by van Mill in [59] that for zero-dimensional
spaces one can show more, as every homogeneous such space is SLH (or, equivalently,
representable [8]). As mentioned in the introduction, all complete SLH-spaces are CDH.
Now this fails for non-complete spaces, e.g. Qω is SLH but not CDH. However, the
connection becomes more interesting if one considers connected spaces (see Problem 1:
van Mill [62] showed that there is a connected, Polish, CDH-space X that is not SLH, while
the problem remains open for continua:
Problem 7. Is every (locally) compact connected separable and metrizable CDH-space SLH?
We do not know to whom this problem belongs. By the result of Kennedy [48], for continua this is equivalent to asking whether every 2-homogeneous metric continuum admits
a homeomorphism that is the identity on some non-empty open set. On the other hand,
there are even Baire SLH connected separable metric spaces which are not CDH [59, 77].
Making use of the celebrated theorem of Effros [20], Ungar [85, 84] (see also [61, 64, 43])
proved a fundamental result on CDH spaces by showing that a sufficiently connected (no
finite set separates) locally compact space X is CDH if and only if it is n-homogeneous
for all n ∈ ω if and only if it is strongly n-homogeneous for all n ∈ ω. van Mill in [67]
extended his result by proving that a connected, CDH-space is n-homogeneous for every n,
and strongly 2-homogeneous when locally connected6. He also presented an example of a
connected CDH-space which is not strongly 2-homogeneous.
Problem 8 (van Mill [67]). Is there a connected separable metric CDH-space which is not
strongly 2-homogeneous?
Moreover, van Mill in [66] showed that Ungar’s theorem works level-by-level, i.e. if a
space X is CDH and no set of size n − 1 separates it, then X is strongly n-homogeneous,
and more importantly, showed that there is a Polish space that is strongly n-homogeneous
for every n ∈ ω which is not CDH.
3.2. Open subsets of CDH spaces. Ungar’s theorem was originally stated as a corollary
to the fact that every open dense subset of a locally compact separable metric CDH-space
is CDH the proof of which is incomplete and remains open:
Problem 9 (Ungar). Is there a locally compact separable metric CDH-space with a dense
open non-CDH-subspace?
In [63] van Mill constructs a Polish space X (a convex subspace of `2 ) with a dense open
rigid subspace, hence showing that there are Polish such spaces.
Several questions considering open subsets of CDH spaces were formulated in [89], one
interesting problem persists:
Problem 10 (Watson-Simon [89]). Are open subsets of CDH continua CDH? (homogeneous)?
6Note
that R is CDH but not strongly 3-homogeneous.
OPEN PROBLEMS
5
3.3. Baire and meager-in-itself CDH-spaces. It was noted by Fitzpatrick and Zhou
in [28] that every CDH-space can be decomposed into two clopen parts, one Baire and
the other meager-in-itself, and that a metric meager-in-itself CDH-space is is a λ-set7.
Hernández-Gutiérrez, Hrušák and van Mill in [40] showed that there is a meager-in-itself
metric space of size κ if and only of there is a λ-set of size κ. In particular, for every
cardinal κ such that ω1 ≤ κ ≤ b there exists a CDH subset of R of size κ. It is consistent
with the continuum arbitrarily large that there is a CDH set of reals of size κ for every
κ < c, while it is also consistent with the continuum arbitrarily large that the only sizes of
CDH metric spaces are ω1 and c. These results were further extended by Medvedev in [57].
Very recently Hrušák and van Mill [43] showed that the existence of a connected CDH
metric space which is meager-in-itself is independent of ZFC. The construction is a nice
combination of techniques of infinite-dimensional topology and set-theory, which, however,
left them wondering about the following:
Problem 11. Is there consistently a connected CDH X ⊆ R2 which is meager-in-itself ?
For more on countable dense homogeneity of subspaces of the plane see [37, 78, 77]
While the possible sizes of meager-in-themselves metric spaces have been determined,
the same is not true for Baire CDH spaces.
Problem 12. Is there consistently a CDH set X ⊆ R of size less than c which is Baire?
It was shown recently in Hernández-Gutiérrez, Hrušák and van Mill [40] that there is
CDH-subset of R of size c which is Baire but not Polish. Consistent examples were before
constructed by Fitzpatrick and Zhou in [28] and by Baldwin and Beaudoin [7]. In fact,
their examples were Bernstein sets8.
Problem 13 (Baldwin and Beaudoin [7]). Is there a Bernstein subset of R which is CDH?
3.4. Products of CDH-spaces and filters. Products of CDH-spaces need not be CDH.
It was first observed by Kuperberg, Kuperberg and Transue in [50] that the square of
the Menger curve M is not even 2-homogeneous, while Bennett [11], building on work of
Anderson [2, 3], showed that M is CDH. A simple example is R × 2ω , which is obviously
not 2-homogeneous, hence not CDH.
The situation changes if one considers only zero-dimensional spaces. In [44] it is shown
that Borel zero-dimensional CDH spaces are productively CDH. Using MA(σ−centered),
Medini in [54] constructed a homogeneous CDH Bernstein set whose square is not CDH.
Problem 14 (Medini [54]). Is there (in ZFC) a separable metric CDH space X such that
X 2 is not CDH?
Hernández-Hrušák-van Mill [40] noted that there is a zero-dimensional CDH X ⊆ R
such that X ω is not CDH. However, the following seems to be open, even omitting the
requirement of zero-dimensionality.
7A
separable metric space X is called a λ-set if every countable subset of X is Gδ in X. A subset X
of a Polish space Y is a λ0 -set if every countable modification of X in Y is a λ-set, i.e. if every countable
subset A of Y is Gδ in X ∪ A.
8A subset X of R is called Bernstein if X and R \ X intersect every Cantor set in R.
6
HRUŠÁK AND VAN MILL
Problem 15 (Medini [54]). For which n ∈ ω is there a separable metric CDH space X
such that X i is CDH for all i ≤ n while X n+1 is not CDH?
On the other hand, there are spaces X which are not CDH while X 2 is (see [58]).
Problem 16 (Medini [54]). For which n ∈ ω is there a separable metric CDH space X
such that X i is not CDH for all i ≤ n while X n+1 is CDH? Can X be a continuum?
Problem 17 (Medini [54]). Is there a zero-dimensional space X that is not CDH while X 2
is CDH? Can X be rigid?
Medini [54] also raises the question whether any of such spaces can be definable (e.g.
analytic or co-analytic)?
The question, whether there is a subset X of R which is not Polish while X ω is CDH,
was raised by Hrušák and Zamora in [44]. The problem was first solved assuming MA by
Medini and Milovich [55] who proved that under MA there is a free ultrafilter p on ω which
is CDH and homeomorphic to its own countable infinite product. This was extended by
Hernández and Hrušák [39] by showing that for every every non-meager P-filter F both
F and F ω are CDH. Finally Kunen, Medini and Zdomskyy [49] showed that the reverse
implication also holds. The following remains a major open problem in combinatorial set
theory:
Problem 18 (see [45]). Is there a non-meager P-filter?
Recently Medini in [53] constructed a non-Polish set of reals X such that X ω is CDH.
Assuming MA+ ω1V = ω1L his example can be made analytic, while ZFC proves that there
are no co-analytic examples and determinacy implies there are no analytic examples. He,
in fact, showed that X ω is CDH for every X ⊆ 2ω such that 2ω \ X is a λ0 - set in 2ω .
While both examples (Medini’s and a non-meager P-filter) are very differently placed
in 2ω - one is possibly homeomorphic to its complement in 2ω , while the complement of
the other contains no perfect set - they both contain many copies of 2ω . In fact, this is a
necessary requrement as Hernández [38] showed that any space X such that X ω is CDH
must contain a copy of the Cantor set. We conclude this discussion by noting that Medini
and Zdomskyy in [56] studied properties which are candidates for characterizing when X ω
is CDH.
Countable dense homogeneity of products of manifolds with boundary was studied by
Yang in [90].
3.5. Baumgartner’s Axiom and κ-CDH. The celebrated theorem of Baumgartner [10]
shows that (assuming the Proper Forcing Axiom PFA) every two ℵ1 -dense9 sets of reals
are order isomorphic. The proof of this influential result in set theory, which was for
instance an important part of Moore’s proof [69] of the consistent existence of a five element
basis for uncountable linear orders, actually proves more: It shows that given two families
{Aα : α < ω1 } and {Bα : α < ω1 } of pairwise disjoint countable dense sets there is a
homeomorphism h ∈ H (R) such that h[Aα ] = Bα for every α < ω1 .
9A
subset A of a space X is κ-dense if |A ∩ U | = κ for every non-empty open U ⊆ X.
OPEN PROBLEMS
7
Inspired by this fact, let us call a separable space X κ-CDH if given two families {Aα : α <
κ} and {Bα : α < κ} of pairwise disjoint countable dense sets there is a homeomorphism
h ∈ H (X) such that h[Aα ] = Bα for every α < κ. Following [83] we denote by BA(X, κ)
the formally weaker property that given two κ-dense subsets A, B of a space X there is a
homeomorphism h ∈ H (X) such that h[A] = B.
It was shown by Abraham and Shelah in [1] that Baumgartner’s result BA(R, ℵ2 ) is
independent of MA. The famous problem of Baumgartner, whether BA(R, ℵ2 ) is consistent,
has attracted a lot of attention recently [70] and apparently has been solved recently in
the affirmative by Neeman [74] who is in the process of writing its very complex proof. It
is very much an open problem, what happens for cardinals above ℵ2 .
Problem 19 (Baumgartner [10]). Is BA(R, ℵ3 ) consistent?
It turns out that dimension 1 is by far the most complicated case here. Steprāns and
Watson in [83] showed that, assuming MAκ (σ−centered) (1) every compact n-dimensional
manifold, for n > 1, is κ-CDH, and consequently (2) Rn for n > 1, is κ-CDH, while Baldwin
and Beaudoin [7] proved that under the same hypothesis 2ω is κ-CDH.
It is not clear, whether there is really any difference between BA(X, κ) and the space X
being κ-CDH. Steprāns and Watson posed the following problems (we phrase them in our
language) which are still open:
Problem 20 (Steprāns and Watson [83]).
(1) Does Rn being κ-CDH imply Rm is κ-CDH for n, m > 1?
(2) Does R being κ-CDH imply Rn is κ-CDH for all n > 1?
(3) Does Rn being κ-CDH imply Rn is λ-CDH for κ > λ?
It can be easily deduced from Ungar’s theorem that every such space is ω-CDH. What
is not clear what happens if the space is not locally compact. The following problems were
asked in a private conversation by Matty Rubin:
Problem 21. Is every Polish CDH space ω-CDH?
Problem 22 (Rubin). Is there a (Polish) space X which is CDH but not 2-CDH?
3.6. Nearly CDH spaces. Hrusak and van Mill in [42] considered the following weakening
of countable dense homogeneity: Call a separable topological space X κ1 -CDH if there are
exactly κ-many types of countable dense subsets of X, i.e. κ-many equivalence classes
of the family of all countable dense sets of X modulo an equivalence relation defined by
D ' D0 if and only if there is an h ∈ H (X) such that h[D] = D0 .
The authors were unaware of an earlier work on the topic by Kennedy [46, 47] who
has much earlier considered the same kind of issues. They would like to use this as an
opportunity to give credit where credit is due. In [46] it is shown that a space which is
1
-CDH for some n ∈ ω has only countably many orbits, all but finitely many of them open,
n
while in [47] the same result is extended to ω1 -CDH spaces, and it is proved there that a
homogeneous continuum with at most countably many '-classes is CDH. These results
were, in retrospect, extended in [42] published more than 30 years later: A locally compact
8
HRUŠÁK AND VAN MILL
separable metric space with at most countably many '-classes has an open dense CDH
subspace whose complement is countable and fixed by all autohomeomorphisms of X. If,
moreover, the space has only n-many '-classes then the complement of the open dense
CDH subspace fixed by all autohomeomorphisms of X has size at most n − 1. It is not
clear whether the same is true for all Polish spaces:
Problem 23. Does every Polish
1
-CDH
ω
space contain a dense open CDH subspace?
Or, say,
Problem 24. Is there a 21 -CDH Polish space without a dense open CDH subspace?
As an extension of a theorem of Hrušák and Zamora-Áviles [44], it is shown in [42] that
every Borel separable metric space which is not Polish is c-CDH. It is also shown in [42]
that every Polish space has either at most countably many '-equivalence classes or c-many
(and for each such possibility there is a locally compact example), or possibly ℵ1 -many. As
of now it is not known whether there is such a Polish (or even locally compact) separable
metric space, a problem equivalent to the Topological Vaught’s Conjecture [86, 72]:
Problem 25. Is there a Polish (locally compact) space which is
1
-CDH?
ω1
The next problem is rather ad hoc. It asks whether there are
essentially different nature than the locally compact ones:
1
-CDH
2
spaces of an
Problem 26. Is there a homogeneous 21 -CDH subset of R?
We conclude this section by including a question of Medini [54]:
Problem 27 (Medini [54]). For which cardinals κ is there a zero-dimensional CDH space
X such that X 2 is κ1 -CDH?
3.7. Countable dense homogeneity with special maps. In the same 1895 volume of
Mathematische Annalen where Cantor published the first result on CDH-spaces - his topological characterization of R - appeared the paper of Stäckel [81] in which it is shown that
given two countable dense subsets of the complex plane C there is an analytic function h
sending one into the other. One of the early results, due to Franklin [34] showed that
Cantor’s argument can be strengthened to show that the witnessing homeomorphism is an
analytic function. The early research concentrated on extensions of Stäckel’s results using
methods of complex analysis (see [89]: Erdős [22] asked if given two countable dense subsets of C there is an entire function sending one onto the other. This was answered in the
affirmative by Maurer [52], while Barth and Schneider [9] extended his result by showing
that such a function can also map the complement of one of the dense sets to the complement of the other. Sato-Rankin [79] and Nienhuys and Thiemann [75] independently
showed that given two countable dense subsets of the real there is an entire transcendental
function (in C) the restriction of which to R is a homemomorphism mapping one of the
dense sets to the other. Dobrowolski [17] and Morayne [71] conluded this line of research by
proving that for any two countable and dense subsets A, B of Cn for n ≥ 2 there exists an
analytic measure preserving diffeomorphism F : Cn → Cn such that F [A] = B. Again the
OPEN PROBLEMS
9
case n = 1 is singular, as the theorem fails there - the only analytic diffeomorphisms from
C onto C are linear. Perhaps the final result in this line of research was provided by Dobrowolski [18] who showed that given a separable Banach space X, a bijection Φ : E → E
with a nowhere dense set of fixed ponts, and A, B countable dense subsets of E there is a
real-analytic diffeomorphism h : E → E such that h−1 ◦ Φ ◦ h[A] = B.
Zamora [91] and Dijkstra [16] studied countable dense homogeneity of metric spaces with
isometries and “almost isometries”, i.e functions such that for some arbitrarily small ε > 0
1−ε<
d(h(x), h(y))
< 1 + ε.
d(x, y)
They call a metric spaces X Lipschitz countable dense homogeneous (Lipschitz-CDH) if
given ε > 0 and A, B countable dense subsets of E there is an ε-almost isometry h such
that h[A] = B. Zamora [91] showed that the Euclidean spaces are Lipschitz-CDH. Dijkstra
[16] extended her result greatly by showing (1) every Lipschitz-SLH space10 is LipschitzCDH and (2) proving that every separable Banach space is Lipschitz-SLH.
Problem 28 (Dijkstra [16]). Is the Hilbert cube with its stadard metric Lipschitz-CDH?
3.8. Non-metrizable CDH spaces. While most of the research on CDH-spaces concentrates on separable metric ones, a fair portion of it deals with general separable spaces.
Arhangel’skii and van Mill in [5] showed among other things that every CDH space has
size at most c and that, assuming CH every compact CDH space is first countable. In the
process, they realized that at the moment no ZFC example of a compact non-metrizable
CDH space was known (there were several consistency results, in particular, it was shown
by Steprāns and Zhou [82] that both 2ω1 and [0, 1]ω1 are CDH under MA+¬CH)
It was shown afterwards by Hernández-Gutiérrez, Hrušák and van Mill [40] that for
every cardinal κ such that there is a λ0 -set of size κ there exists a compact CDH-space of
weight κ, hence, in particular, there is a compact CDH-space of weight ω1 in ZFC. This,
however, leaves the following problem open:
Problem 29 (Hernández-Hrušák-van Mill [40]). Is there (in ZFC) a compact CDH space
of weight c?
Arhangel’skii and van Mill in [5] showed that the double arrow space is not CDH. It
was believed at the time that a trivial modification of it may be CDH. These hopes were
trashed by Hernández [38] who showed that none of the trivial modifications is CDH, and
also that the double arrow space itself is quite far from being CDH, as it is in fact c-CDH.
Inspired by his results we ask:
Problem 30. Is there a compact CDH-space which does not contain a copy of the Cantor
set?
Also the space constructed in [40] is zero-dimensional.
10Lipschitz-SLH
isometries.
is the obvious modification of SLH, requiring that the corresponding maps be almost
10
HRUŠÁK AND VAN MILL
Problem 31 (Hernández-Hrušák-van Mill [40]). Is there a non-metrizable CDH continuum
in ZFC?
Steprāns and Zhou [82] studied countable dense homogeneity of separable non-metrizable
manifolds. In particular, they showed that any separable manifold of weight < b is CDH,
while there is a non-CDH separable manifold of weight c. Watson in [87] showed that
consistently there is a separable, non-CDH manifold of weight < c. The following question
is a bold reformulation of a question of Watson in [88]:
Problem 32. Is b the minimal weight of a separable, non-CDH manifold?
If not, is there (as Watson asks) a nice combinatorial characterization of the minimal
weight of a separable, non-CDH manifold?
References
[1] U. Avraham and S. Shelah, Martin’s axiom does not imply that every two ℵ1 -dense sets of reals are
isomorphic, Israel J. Math. 38 (1981), no. 1-2, 161–176.
[2] R. D. Anderson, A characterization of the universal curve and a proof of its homogeneity, Ann. of
Math. 67 (1958) 313–324.
[3] R. D. Anderson, One dimensional continuous curves and a homogeneity theorem, Ann. of Math. 68
(1958) 1–16.
[4] R. D. Anderson, D. W. Curtis and J. van Mill, A fake topological Hilbert space, Trans. Amer. Math.
Soc. 272, (1982), 311–321.
[5] A. V. Arhangel’skii and J. van Mill, On the cardinality of countable dense homogeneous spaces, Proc.
Amer. Math. Soc. 141, (2013), 4031–4038.
[6] A. V. Arhangel’skii and J. van Mill, Topological homogeneity, in Recent Progress in General Topology
III. (K. P. Hart, J. van Mill, and P. Simon, eds.), Atlantis Press, Paris, 2014, 1–68.
[7] S. Baldwin, R. E. Beaudoin, Countable dense homogeneous spaces under Martin’s axiom, Israel J.
Math. 65 no. 2, (1989), 153–164.
[8] J. W. Bales, Representable and strongly locally homogeneous spaces and strongly n-homogeneous
spaces, Houston J. Math. 2 (1976), no. 3, 315–327.
[9] K. F. Barth, W. J. Schneider, Entire functions mapping arbitrary countable dense sets and their
complements onto each other, J. London Math. Soc. 4 (1971/72), 482–488.
[10] J. E. Baumgartner, Applications of the proper forcing axiom, in: K. Kunen and J.E. Vaughan, eds.,
Handbook of Set-Theoretic Topology (North-Holland, Amsterdam, 1984), 913–955.
[11] R. B. Bennett, Countable dense homogeneous spaces, Fund. Math. 74 (1972) 189–194.
[12] C. Bessaga and A. Pelczyński, The estimated extension theorem homogeneous collections and skeletons,
and their application to the topological classification of linear metric spaces and convex sets, Fund.
Math. 69 (1970), 153–190.
[13] L. E. J. Brouwer, Some remarks on the coherence type η, Proc. Akad. Amsterdam 15 (1913), 12561263.
[14] C. E. Burgess, Some theorems on n-homogeneous continua, Proc. Amer. Math. Soc. 5, (1954) 136–143.
[15] G. Cantor, Baiträge zur Begründung der transfiniten Mengenlehre, Math. Ann. 46 (1895) 481–512.
[16] J. J. Dijkstra, Homogeneity properties with isometries and Lipschitz functions, Rocky Mountain J.
Math. 40 (2010), no. 5, 1505–1525.
[17] T. Dobrowolski, On smooth countable dense homogeneity, Bull. Acad. Polon. Sci. Sér. Sci. Math.
Astronom. Phys. 24 (1976), 627–634.
[18] T. Dobrowolski, A conjugated version of the countable dense homogeneity, Topology Appl. 156 (2009),
no. 4, 795–802.
OPEN PROBLEMS
11
[19] A. Dow and E. Pearl, Homogeneity in powers of zero-dimensional first-countable spaces, Proc. Amer.
Math. Soc. 125 (1997), 2503–2510.
[20] E. G. Effros, Transformation groups and C ∗ -algebras, Ann. of Math. 81 (1965), 38–55.
[21] F. van Engelen, On the homogeneity of infinite products, Topology Proc. 17 (1992), 303–315.
[22] P. Erdős, Some unsolved problems, Michigan Math. J. 4 (1957), 291–300.
[23] I. Farah, M. Hrušák, and C. Martı́nez Ranero, A countable dense homogeneous set of reals of size ℵ1 ,
Fund. Math. 186 (2005), 71–77.
[24] B. Fitzpatrick, Jr., A note on countable dense homogeneity, Fund. Math. 75 (1972), 33–34.
[25] B. Fitzpatrick, Jr. and N. F. Lauer, Densely homogeneous spaces. I, Houston J. Math. 13 (1987),
19–25.
[26] B. Fitzpatrick, Jr. and H.-X. Zhou, Densely homogeneous spaces. II, Houston J. Math. 14 (1988),
57–68.
[27] B. Fitzpatrick, Jr. and H. X. Zhou, Some open problems in densely homogeneous spaces, Open Problems in Topology (J. van Mill and G. M. Reed, eds.), North-Holland Publishing Co., Amsterdam,
1990, pp. 252–259.
[28] B. Fitzpatrick, Jr. and H.-X. Zhou, Countable dense homogeneity and the Baire property, Topology
Appl. 43 (1992), 1–14.
[29] B. Fitzpatrick, J. M. S. White, and H.-X. Zhou, Homogeneity and σ-discrete sets, Proceedings of the
Symposium on General Topology and Applications (Oxford, 1989). Topology Appl. 44 (1992), no.
1-3, 143–147.
[30] B. Fitzpatrick, Jr. and H.-X. Zhou, A survey of some homogeneity properties in topology, Papers on
general topology and related category theory and topological algebra (New York, 1985/1987), 28–35,
Ann. New York Acad. Sci., 552, New York Acad. Sci., New York, 1989.
[31] P. Fletcher. and R. A. McCoy, Conditions under which a connected representable space is locally
connected, Pacific J. Math. 51, (1974) 433–437.
[32] L. R. Ford, Jr., Homeomorphism groups and coset spaces , Trans. Amer. Math. Soc. 77 (1954), 490–
497.
[33] M. K. Fort, Jr., Homogeneity of infinite products of manifolds with boundary, Pacific J. Math. 12
(1962), 879–884.
[34] P. Franklin, Analytic transformations of everywhere dense point sets, Trans. Amer. Math. Soc. 27
(1925), 91–100.
[35] M. Fréchet, Les dimensions dun ensemble abstrait, Math. Ann. 68 (1910), 145–168.
[36] J. de Groot, Topological Hilbert space and the drop-out effect, Rapport ZW 1969-016, Mathematisch
Centrum, Amsterdam, 1969.
[37] W. R. Heath and W. L. Saltsman, A Jones space that is not countable dense homogeneous, Topology
Appl. 52 (1993), no. 3, 237–240.
[38] R. Hernández-Gutiérrez, Countable dense homogeneity and the double arrow space, Topology Appl.
160 (2013), no. 10, 1123–1128.
[39] R. Hernández-Gutiérrez, M. Hrušák, Non-meager P-filters are countable dense homogeneous, Colloq.
Math. 130 (2013), 281–289.
[40] R. Hernández-Gutiérrez, M. Hrušák, and J. van Mill, Countable dense homogeneity and λ-sets, Fund.
Math. 226 (2014), 157–172.
[41] R. Hernández-Gutiérrez and P. Szeptycki, Some observations on filters with properties defined by open
covers, Comment. Math. Univ. Carolin. 56, (2015), no.3, 355–364.
[42] M. Hrušák, and J. van Mill, Nearly countable dense homogeneous spaces, Canad. J. Math. 66 (2014),
743–758.
[43] M. Hrušák and J. van Mill, The existence of a connected meager in itself CDH space is independent
of ZFC, Proc. Amer. Math. Soc., to appear.
12
HRUŠÁK AND VAN MILL
[44] M. Hrušák, and B. Zamora-Áviles, Countable dense homogeneity of definable spaces, Proc. Amer.
Math. Soc. 133 (2005), 3429–3435.
[45] W. Just, A. R. D. Mathias, K. Prikry, and P. Simon, On the existence of large P-ideals, J. Symbolic
Logic 55 (1990), 457–465.
[46] J. Kennedy, A weaker form of countable dense homogeneity, Houston J. Math. 3 no. 1 (1977), 89–97.
[47] J. Kennedy, More on a weaker form of countable dense homogeneity, Houston J. Math. 10 (1984), no
4, 507–524.
[48] J. Kennedy, A condition under which 2-homogeneity and representability are the same in continua,
Fund. Math. 121 (1984), 89–98.
[49] K. Kunen, A. Medini and L. Zdomskyy, Seven characterizations of non-meager P-filters, Fund. Math
231 (2015), no. 2, 189–2018.
[50] K. Kuperberg, W. Kuperberg and W. R. R. Transue, On the 2-homogeneity of Cartesian products,
Fund. Math 110 (1980), no. 2, 131–134.
[51] L. B. Lawrence, Homogeneity in powers of subspaces of the real line, Trans. Amer. Math. Soc 350
(1984), no. 8, 3055–3064.
[52] W. D. Maurer, Conformal equivalence of countable dense sets, Proc. Amer. Math. Soc. 18 (1967),
269–270.
[53] A. Medini, Countable dense homogeneity in powers of zero-dimensional definable spaces, Canad. Math.
Bull. 58 (2015), no.2, 334–349.
[54] A. Medini, Products and countable dense homogeneity, Top. Proc. 46 (2015), 135-143.
[55] A. Medini and D. Milovich, The topology of ultrafilters as subspaces of 2ω , Topology Appl. 159 (2012),
no. 5, 1318–1333.
[56] A. Medini and L. Zdomskyy, Between Polish and completely Baire, Arch. for Math. Logic 54 (2015),
no. 1-2, 231–245.
[57] S. V. Medvedev, Metrizable DH-spaces of the first category, Topology Appl. 179 (2015), 171–178.
[58] J. van Mill, A rigid space X for which X × X is homogeneous; an application of infinite-dimensional
topology, Proc. Amer. Math. Soc. 83 (1981), no.3, 597–600.
[59] J. van Mill, Strong local homogeneity does not imply countable dense homogeneity, Proc. Amer. Math.
Soc. 84 (1982), 143–148.
[60] J. van Mill, The infinite-dimensional topology of function spaces, North-Holland Publishing Co., Amsterdam, 2001.
[61] J. van Mill, A note on the Effros Theorem, Amer. Math. Monthly. 111 (2004), 801–806.
[62] J. van Mill, On countable dense and strong local homogeneity, Bull. Pol. Acad. Sci. Math. 53 (2005),
401–408.
[63] J. van Mill, A countable dense homogeneous space with a dense rigid open subspace, Fund. Math. 201
(2008), 91–98.
[64] J. van Mill, Homogeneous spaces and transitive actions by Polish groups, Israel J. Math. 165 (2008),
133–159.
[65] J. van Mill, Analytic groups and pushing small sets apart, Trans. Amer. Math. Soc. 361 (2009), no.
10, 5417–5434.
[66] J. van Mill, On countable dense and strong n-homogeneity, Fund. Math. 214 (2011), no. 3, 215–239.
[67] J. van Mill, On countable dense and n-homogeneity, Can. Math. Bull. 56 (2013), 860–869.
[68] J. van Mill, Countable dense homogeneous rimcompact spaces and local connectivity, Filomat 29
(2015), no. 1, 179–182.
[69] J. T. Moore, A five element basis for the uncountable linear orders, Ann. of Math. (2), 163 (2006),
no. 2, 669–688.
[70] J. T. Moore and S. Todorcevic, Baumgartner’s isomorphism problem for ℵ2 -dense suborders of R,
Archive for Math. Logic, to appear.
[71] M. Morayne, Measure preserving analytic diffeomorphisms of countable dense sets in Cn and Rn ,
Colloq. Math. 52 (1987), 93–98.
OPEN PROBLEMS
13
[72] M. Morley, The number of countable models, J. Symbolic Logic 35 (1970), 14–18.
[73] D. B. Motorov, Homogeneity and π-networks, Moscow Univ. Math. Bull. 44 (1989), no.4, 45–50.
[74] I. Neeman, http://settheory.mathtalks.org/appalachian-set-theory-workshop-itay-neeman-april-92016/
[75] J. W. Nienhuys and J. G. F. Thiemann, On the existence of entire functions mapping countable dense
sets onto each other, Indag. Math. 38 (1976), 331-334.
[76] D. Repovš, L. Zdomskyy and S. Zhang, Countable dense homogeneous filters and the Menger covering
property, Fund. Math 224 (2014), no. 3, 233–240.
[77] W. L. Saltsman, Concerning the existence of a connected, countable dense homogeneous subset of the
plane which is not strongly locally homogeneous, Topology Proc. 16 (1991), 137–176.
[78] W. L. Saltsman, Concerning the existence of a nondegenerate connected, countable dense homogeneous
subset of the plane which has a rigid open subset, Topology Proc. 16 (1991), 177–183.
[79] D. Sato and S. Rankin, Entire functions mapping countable dense subsets of the reals onto each other
monotonically, Bull. Austral. Math. Soc. 10 (1974), 67-70.
[80] W. Sierpiński, Sur une propriété topologique des ensembles d enombrables denses en soi, Fund. Math.
1 (1920), 11–16.
[81] P. Stäckel, Ueber aritmetishe Eigenschaften analytischer Functionen, Math. Ann. 46 (1895) 513–520.
[82] J. Steprāns and H. X. Zhou, Some results on CDH spaces. I, Topology Appl. 28 (1988), 147–154.
[83] J. Steprāns and S. Watson, Homeomorphisms of manifolds with prescribed behaviour on large dense
sets, Bull. London Math. Soc., 19 (1987), 305–310.
[84] G. S. Ungar, On all kinds of homogeneous spaces, Trans. Amer. Math. Soc. 212 (1975), 393–400.
[85] G. S. Ungar, Countable dense homogeneity and n-homogeneity, Fund. Math. 99 (1978), 155–160.
[86] R. L. Vaught, Denumerable models of complete theories, In: Infinitistic Methods (Proc. Sympos.
Foundations of Math., Warsaw,1959), Pergamon, Oxford; Państwowe Wydawnictwo Naukowe, Warsaw, 1961, 303–321.
[87] S. Watson, The homogeneity of small manifolds, Topology Proc. 13 (1988), no. 2, 365–369.
[88] S. Watson, Problems I wish I could Solve, Open Problems in Topology (J. van Mill and G. M. Reed,
eds.), North-Holland Publishing Co., Amsterdam, 1990.
1990, pp. 38–76.
[89] S. Watson and P. Simon, Open subspaces of countable dense homogeneous spaces, Fund. Math. 141
(1992), 101–108.
[90] Z. Yang, Homogeneity of infinite products of manifolds with boundary, Papers on general topology
and applications (Brookville, NY, 1990), 194–208, Ann. New York Acad. Sci., 659, New York Acad.
Sci., New York, 1992.
[91] B. Zamora-Áviles, Espacios numerablemente densos homogéneos, Master’s Thesis (in Spanish), UMSNH, Mexico, (2003).
[92] H.-X. Zhou, Two applications of set theory to homogeneity, Questions Answers Gen. Topology 6
(1988), no. 1, 49–56.
1
Centro de Ciencas Matemáticas, UNAM, A.P. 61-3, Xangari, Morelia, Michoacán,
58089, México
E-mail address: [email protected]
KdV Institute for Mathematics, University of Amsterdam, Science Park 904, P.O. Box
94248, 1090 GE Amsterdam, The Netherlands
E-mail address: [email protected]